Language-based event-related potentials in children with and without specific language impairment

This study is one of the first investigations using event-related potentials (ERPs) to measure the neural correlates of semantic and syntactic processing in children with specific language impairment (SLI). These children have difficulty with both semantic and syntactic processing, but show a particular syntactic deficit in the use of finiteness markers. Although the linguistic deficits of children with SLI have been extensively studied using behavioral measures, little is known about the neural correlates of these behavioral disorders.; Two ERP components were measured: the semantic N400 and the syntactic P600. Of primary interest was the presence of these components in children with SLI. However, because the nature of these components is largely unknown in typically developing children, the ERPs of control children were also investigated. The participants included one adult group and four child groups varying in age (young and old) and affectedness (SLI and control). ERPs were recorded while participants listened to sentences containing a semantic anomaly, a syntactic error, or no error. After each stimulus participants judged the grammaticality of the item.; The grammaticality judgment results showed that the young SLI group had difficulty detecting all error types whereas the other groups performed near ceiling level on all stimulus types. The ERP results revealed the adult group showed the expected components: an N400 at fronto-central and centro-parietal midline and right hemisphere sites and a P600 following overt and omitted finiteness errors at centro-parietal sites. The control children showed fewer ERP components than expected. The old control children showed a fronto-central/centro-parietal N400 and a fronto-central/centro-parietal P600 following omitted finiteness errors. No significant ERPs were found for the young control group. Compared to the control groups, the old SLI group showed a larger and more broadly distributed P600 effect following omitted finiteness errors. The young SLI group did not demonstrate a P600. These P600 results indicated that for children with SLI whose grammatical deficits may have recently resolved, the related ERP component is larger, perhaps suggesting a heightened awareness of syntactic errors. However, for children with SLI who showed behavioral deficits, no significant ERP components were found.